In single-chip color cameras, images captured by a single exposure are separated into several, non-overlapping, two-dimensional patterns encoded to component colors. Only one component color of the image is captured per pixel. Color filters must be patterned to cover individual pixels to permit suitable sampling of spatial color.
Recently, a photo-crosslinkable red-green-blue (RGB) Bayer color filter pattern is being widely used for solid-state sensor and display. As the pixel size decreases, color distortion caused by the misalignment of different color patterns becomes serious and the green pattern (photo-crosslinkable polymer) needs to be exposed twice. Additionally, the differences in thickness of the three spin-coated color (RGB) filters cause different photo-transmission rates and needs a polymer spin coat planarization layer for the following microlens (μ-lens) process.
In the conventional color filter and μ-lens process for pixel sizes smaller than about 4 μm, there are six (6) masking layers: blue, green-1, green-2, red, an overlying planarization layer to accommodate the difference thicknesses of the previous four color masking layers and the μ-lens layer.
U.S. Pat. No. 6,258,065 B1 to Levy discloses a coplanar top surface of color filters using an etch back of the third color layer.
U.S. Pat. No. 5,899,277 to Hawkins et al. describes coplanar top and bottom surfaces of color filters.
U.S. Pat. No. 4,315,978 to Hartman describes uneven color filters.
U.S. Pat. No. 5,719,074 to Hawkins et al. describes a method of making a planar color filter array for CCDS from dyed and mordant layers.